As computer technologies continue to push the bounds of ever increasing performance and innovation, computer system hardware and software requirements have dramatically increased in complexity and sophistication. This has become even more apparent with the growth of network technologies such as the Internet, for example, wherein ever increasing amounts of data are continually being operated upon, requested and transported. Other applications such as database managers are also being pushed to handle this increasing traffic of data. One technology that has emerged to transport and process data is XML, which is often referred to as a meta-markup language for text documents. Data can be described in XML documents as strings of text, wherein the actual data is encapsulated by text markup that describes the data. A unit of data and associated markup is defined as an element. Currently, XML defines how elements are delimited by tags, what a tag looks like, what names are acceptable for elements, and where attributes are placed, among other specifications, for example.
The markup in an XML document describes the underlying structure of the document and generally defines which elements are associated with other elements. The markup can also include document semantics. As an example, the markup can indicate that an element is a date, serial number, name or other type of code or designation. As these and other types of elements are transmitted (e.g., XML stream) between systems or databases, changes are often required to be made to existing data due to recently discovered errors in the data or due to the data having changed over time (e.g., person's name or address changing). Since XML merely indicates the most recent state of data, problems can occur as data is transported between databases and/or systems.
One such problem relates to errors having been previously transmitted to a database via an XML data stream. As an example, if a data record containing a person's name were erroneously transmitted with an incorrect middle initial, simply re-transmitting the name with the correct initial would not necessarily correct the previous error. For example, a new database entry with a different name could exist. Unless the erroneous name or data record were somehow purged from the database—which can cause an expensive amount of database processing, the database can become more difficult to administer over time and can grow past designed system memory boundaries. Similar problems exist when data records need to be changed or altered.